Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add instructions/sam-agents/sam/agents-mdgit clone --depth 1 https://github.com/sam-agents/samWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.01871 | $0.01871 |
| Opus 5 | $0.00936 | $0.00936 |
| Sonnet 5 | $0.00374 | $0.00374 |
| Haiku 4.5 | $0.00187 | $0.00187 |
Grade B, and why
sam AGENTS.md scanned grade B with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Enumerates other installed skillsmediumAgent snooping
Other skills' SKILL.md files reveal prompts, capabilities and secrets that should be invisible to peers.
ls -la ./test-project/.gemini/skills/ # for Gemini CLI How it starts
The opening of the file, as written. The whole thing — 297 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md - Agent Coding Guidelines for SAM
This file provides guidelines for agentic coding agents operating in this repository.
Project Overview
SAM (Smart Agent Manager) is an autonomous TDD agent system that generates agent configurations for Claude Code, Cursor IDE, Antigravity, and Gemini CLI. The project consists of:
- CLI (
bin/cli.js) - Node.js CLI tool that installs agent configurations - Templates (
templates/) - Agent definitions in markdown format - Shared configs (
_sam/) - Agent definitions used across all platforms
Build / Lint / Test Commands
Testing the CLI
# Test CLI locally (interactive mode)
node bin/cli.js
# Test CLI with specific platform
node bin/cli.js --platform claude ./test-project
node bin/cli.js --platform cursor ./test-project
node bin/cli.js --platform antigravity ./test-project
node bin/cli.js --platform gemini ./test-project
node bin/cli.js --platform all ./test-project
# Test with target directory
node bin/cli.js ./my-project
node bin/cli.js --platform cursor ./my-project
Running a Single Test
Formal test suites are available for specific platforms:
- Gemini CLI:
npm testrunsscripts/verify-gemini.jsto validate skill generation.
Manual testing is done by:
- Creating a test directory:
mkdir test-project - Running the CLI:
node bin/cli.js --platform <platform> ./test-project - Verifying output files:
ls -la ./test-project/_sam ls -la ./test-project/.claude/commands/sam # for Claude Code ls -la ./test-project/.cursor/rules/ # for Cursor ls -la ./test-project/.agent/skills/ # for Antigravity ls -la ./test-project/.gemini/skills/ # for Gemini CLI
Publishing
# Publish to npm (requires authentication)
npm publish --access public --provenance
Code Style Guidelines
JavaScript (CLI)
Follow these conventions in bin/cli.js:
- Formatting
- 2-space indentation
- No semicolons at line endings
- Single quotes for strings
- Trailing commas in multi-line objects/arrays
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 297 lines · 1,871 tokens per session scan B efb79ecd9255
sam AGENTS.md is an instructions file published in the GitHub repository sam-agents/sam (18 stars, last pushed 1mo ago), licensed MIT. It adds 1,871 tokens to every session, about $0.0094 per session on Opus 5. A static security scan graded it B with 1 finding (enumerates other installed skills). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
aiworkspace AGENTS.md
Instructions for a-tokyo/aiworkspace, covering ai workspace, structure, key commands, conventions and code quality.
agents CLAUDE.md
Claude Code instructions for wshobson/agents, a project described as: Multi-harness agentic plugin marketplace for Claude Code, Codex, Cursor, OpenCode, GitHub Copilot, and Google Antigravity.
agent-skills CLAUDE.md
Instructions for LLl0k0laD/agent-skills, covering swe-skills, project structure, skills by phase, conventions and commands.
ai-workspace-archive copilot-instructions.md
Copilot instructions for HIDORAKAI002/ai-workspace-archive, covering github copilot instructions for autogpt, repository overview, build and validation instructions, essential setup commands and runtime requirements.
capy CLAUDE.md
Claude Code instructions for serpro69/capy, covering claude.md, project, architecture, key packages and critical invariants.
awesome-agentic-AI-coding-template AGENTS.md
Instructions for florian101010/awesome-agentic-AI-coding-template, covering agents.md, project overview, architecture (must-know), file structure and coding rules.